Should You Sell a Large IPv4 Block Whole or Split It?

Anyone who decides to sell a large IPv4 block soon faces the same question: offer the whole range to one buyer, or cut it into smaller prefixes? A whole-block sale is simpler and ends with a single transfer. Smaller blocks reach more buyers and can sell at a higher price per address, but that only pays off if enough of them actually sell.

Price per address is the number everyone quotes. The number that decides the outcome is net proceeds by a date the seller can live with. The rest of this article works through that comparison with a worked example.

Why Large IPv4 Blocks Trade Differently

A /16 contains 65,536 addresses. At $15 per address that is $983,040 in a single payment, which only a few telecom operators, cloud and hosting companies and infrastructure investors can commit at the same time.

A /19 (8,192 addresses) or a /21 (2,048) costs a fraction of that and fits the budget of many more companies. Buyers for a /21 are far easier to find than buyers for a /16, and that liquidity is what the higher price per address pays for. The size of the premium is not fixed: it widens and narrows with market conditions.

Sell the Entire IPv4 Block or Split It?

How Block Size Changes the Buyer Pool

One /16 can be offered as:

• two /17 blocks;
• four /18 blocks;
• eight /19 blocks;
• sixteen /20 blocks;
• thirty-two /21 blocks;
• or 256 individual /24 blocks.

The address count is the same in every case, but the buyers are not. A company shopping for a /21 has a different budget, use case and approval process from one considering a /16. Our guide to IPv4 prefix length covers how CIDR size affects address count, routing and planning.

More buyers is not automatically better. Cutting a /16 into 256 /24s would mean hundreds of separate negotiations, checks and transfers. In practice the split follows demand: the seller picks the sizes buyers are asking for now, not the smallest prefix the registry will accept.

Does Splitting an IPv4 Block Increase the Net Sale Price?

Take a hypothetical /16 and two ways to sell it:

• the complete /16 at $15 per address;
• eight /19s at $17 per address.

If all eight /19s sell, gross proceeds are $131,072 higher than in the whole-block sale. If only seven sell, they are $8,192 lower, and the seller still holds one /19. Transaction costs and the wait for that last sale push the result further against the split.

The break-even point is the ratio of the two prices:

share of addresses that must sell = whole-block price per IP ÷ split price per IP

Here that is $15 ÷ $17, or about 88.24%. Seven of eight /19s is 87.5%, which falls short, so in this example every /19 has to sell before the split comes out ahead. The test is on gross cash and counts an unsold remainder as worth nothing during the period. A real remainder keeps some value, but it can only be sold later and at the cost of another transaction.

The prices are for illustration only. Actual pricing depends on the RIR, block status, size, reputation, routing history, geolocation, payment terms and current demand.

What Costs Change When a Block Is Split?

Commission scales with the amount sold, so it is easy to model. Other costs repeat with every buyer: each one needs its own contract, compliance review, escrow, invoice and RIR request, and some fees are charged per transaction rather than per address.

Put these into the model:

• brokerage or marketplace commission;
• escrow and payment-processing fees;
• registry, LIR or sponsoring-LIR charges where they apply (the RIPE NCC states that its transfers are free of charge, while other registries and sponsoring LIRs may bill);
• legal and compliance work for each buyer;
• the seller’s own time spent approving and signing several deals;
• currency risk if sales close over several months;
• the cost of receiving capital later instead of at one closing.

Compare the two strategies on expected net proceeds, not on the advertised price per IP.

The Risk of Selling Only the Best Subnets

Not every part of a block sells equally easily. Buyers look at where a prefix starts and ends, at its routing history, and at whether it has been associated with an unwanted hostname, network or GeoIP profile. If the best parts go first, the rest may have to be discounted later.

Fragmentation matters too. A complete /17 can interest a larger operator, while a few non-contiguous /20s with gaps between them sell into a different market. Before accepting the first partial offer, work out exactly what would be left and whether it still forms useful CIDR blocks.

Check Transfer Eligibility Before Advertising Smaller Blocks

A block can be split on paper and still not be transferable in that form. Before promising any buyer a particular prefix, check the registered holder, the resource status, the RIR’s policy and how long the seller has held the block. These checks belong in the wider IPv4 transfer process.

The main rules differ by registry:

• ARIN: the minimum IPv4 transfer size is a /24. To qualify for a larger initial or additional block, the recipient must document the use of at least 50% of the requested space within 24 months. Recipients that already hold IPv4 resources are also subject to utilisation requirements for their existing holdings.
• APNIC: the minimum transfer size is also a /24. Addresses delegated from the 103/8 free pool cannot be transferred for five years after the original delegation. A recipient in the APNIC region must provide a detailed plan for using the transferred resource within 24 months; existing resource holders must also show their past usage and compliance with APNIC policies.
• RIPE NCC: IPv4 addresses and 16-bit ASNs cannot be transferred for 24 months after they were received from the RIPE NCC, through a transfer from another organisation or following a change in business structure such as a merger or acquisition. Complete or partial IPv4 blocks can be transferred.

The fact that ARIN and APNIC accept transfers as small as a /24 does not make that degree of splitting commercially sensible. Agree the sale plan only after the registration history and the requirements of both source and recipient have been reviewed.

Prepare Routing and Registry Data for Each Prefix

Each prefix that sells needs a clean handover. Depending on the deal and how the block is used, the parties may need to update or remove:

• RDAP and WHOIS registration data;
• IRR route objects;
• RPKI ROAs;
• reverse-DNS delegations;
• legacy LOAs and old routing authorisations;
• previous BGP announcements;
• geofeed and GeoIP correction records.

With one buyer this is done once. With eight or thirty-two, the seller needs a fixed order of work so that moving one prefix does not disrupt the ones still for sale.

When Selling the Complete Block Makes Sense

A whole-block sale is usually the stronger option when:

• speed and certainty matter more than the last dollar per address;
• a qualified buyer is ready to take the entire resource;
• the price gap between whole-block and smaller-block offers is narrow;
• you want one contract, one escrow process and one RIR transfer;
• management does not want to keep leftover IPv4 inventory;
• part of the block’s value lies in its size, continuity or routing history.

A lower price per address can still be the better deal if it turns the whole block into cash months sooner and leaves nothing unsold.

When Splitting the Block Is the Better Strategy

Splitting deserves serious consideration when:

• current demand is concentrated in /19, /20 or /21 blocks;
• the expected price premium clearly exceeds the extra costs;
• you can live with a staged sale;
• the block divides into clean, useful and transferable prefixes;
• several credible buyers are already available;
• you can hold and manage the remainder without pressure to sell.

A staged approach can lower the risk: keep a larger contiguous portion in reserve and test demand with one or two smaller prefixes. That only works if the boundaries are planned before the first sale.

Decision Checklist for IPv4 Sellers

Before choosing a structure, answer these questions:

1. What is the realistic whole-block price, as opposed to the initial asking price?
2. Which smaller sizes have active buyers right now?
3. How many portions have to sell for the split to break even?
4. What will the remaining CIDR inventory look like after each partial sale?
5. How long could the complete disposal take?
6. Which costs repeat for every buyer?
7. Is each proposed prefix eligible for transfer?
8. Are registry, routing, RPKI, IRR and reverse-DNS records ready for a staged handover?
9. Can the organisation keep managing an unsold remainder?

Compare Net Proceeds, Not Price per IP

The two approaches have different economics. A whole-block sale concentrates the search on one buyer and requires only one closing. A split reaches a wider market but creates a series of transactions whose final result is less predictable.

Compare them on expected net proceeds over a fixed period. The calculation should cover the chance that every portion sells, what the remainder looks like and the repeated transaction costs. If the split still comes out ahead after that, the higher price per address is real.